2026-10-13T12:24:08Z https://api.figshare.com/v2/oai

oai:figshare.com:article/33872278 2026-09-17T05:58:38Z category_1730 portal_316 item_type_3 month_year_09_2026

Data Sheet 1_Automated detection of mutual gaze in mobile eye-tracking studies of multi-party interactions.zip Danny Dirker (25002364) Anja Stukenbrock (9368642) Communication Studies automated gaze detection computer vision eye-tracking multi-party interaction mutual gaze <p>This article describes a data-processing pipeline for quantifying mutual gaze in mobile eye-tracking studies. Mutual gaze is defined as eye contact between participants who look at each other simultaneously. We applied the pipeline to data from dyadic and multi-party interactions (“quartets”). Using state-of-the-art computer vision tools for fiducial-marker tracking and face detection, the pipeline processes scene camera videos from TobiiPro eye-trackers to extract facial areas-of-interest (AOIs). Each participant's gaze samples are classified relative to visible co-participants, indicating when gaze is directed at another's face. By integrating multiple participants' high-resolution data streams, the pipeline identifies periods of unreciprocated (unilateral) gaze as well as mutual gaze. The output files contain time-stamped measures of gaze events and can be imported into the annotation tool ELAN for qualitative analysis. While processing is largely automated, some steps require validity checks by the researcher. To underpin our observations on the potential of the proposed pipeline for the automatic extraction of mutual gaze, we subsequently conduct an exploratory analysis on the participants' unilateral gaze and mutual gaze proportions in relation to the following factors: seating position, speaking proportion, age, gender, familiarity, and movement synchrony. The results show that speaking proportion, seating position, and familiarity were significantly associated with unilateral and mutual gaze proportion. By contrast, an association with age, gender, and movement synchrony measurements was not shown. The article concludes with a critical discussion of the limitations and strengths of the proposed pipeline. We highlight the opportunities and challenges of quantifying gaze patterns in complex multi-party settings with a specific interest on how existing qualitative and quantitative findings may be brought together for future studies.</p> 2026-09-17T05:58:38Z Dataset Dataset 10.3389/fcomm.2026.1857066.s002 https://figshare.com/articles/dataset/Data_Sheet_1_Automated_detection_of_mutual_gaze_in_mobile_eye-tracking_studies_of_multi-party_interactions_zip/33872278 CC BY 4.0